Controlling water in underground coal mines is an important safety consideration. There are very few methods that can detect or map water remotely within the working face of a mine. The two-dimensional (2D) electrical resistivity tomography (ERT) method is sufficient for detecting water directly above the roof and below the floor, but it fails to detect water within the internal blocks of the working face. Therefore, it is critical to develop a new acquisition methodology for mapping internally within the face given the limited space in which to work. In this paper, we apply the pole-pole array in an L-array around the working face using a modified roll-along method. The data are then inverted using commercial software. With our new methodology, abnormal areas that likely indicate excess water are delineated to provide guidance for the control of mine water.